Classification-Based Management of Stiff/Ankylosed Knees

Abstract

Introduction

Stiff knees are defined as those with an arc of motion < 50°. They pose a considerable surgical challenge to the operating surgeon. Based on our experience to deal with these complex cases, we have developed a working classification that outlines a flowchart to manage the stiff/ankylosed knees.

Materials and Methods

It was a retrospective study conducted in our department. Out of 570 TKA performed in last 5 years, 57 had stiffness and four had bony ankyloses (total 61 knees). Patients were classified based on the fibrous or bony ankylosis and preoperative ROM.

Results

Patients were followed for an average 2.4 years (1.8–5.5 years). KSS pain scores improved from an av. 32 preop (18–64) to av. 76 postoperatively (61–90). The KSS function scores improved from a preoperative value of 36 (16–56) to an av. 78 (52–90) postoperatively. ROM improved from an average of 35.6° (0°–44°) preoperatively to an average of 95.6° (ROM 73°–118°) postoperatively. Extension lag was an av. 8° (3°–12°) and was seen in 13 patients postoperatively. Residual fixed flexion deformity was an av. 7° (3°–14°) and seen in 17 patients. The stiff knees (type 1 and type 2) fared better than ankylosed knees (type 3) in all aspects. Complication rate was high (24%) in our series.

Conclusion

Our classification of stiff/ankylosed knees guides the surgeon to decide upon which approach to take, which implants to keep handy and has a predictive and prognostic value.

Introduction

Stiffness has been defined as an arc of motion of < 50° [1]. Normal activities of daily living require a certain degree of range of motion (ROM) at the knee joint for adequate function. Normal walking requires 70°–80° ROM, stairs require 80°–90° of ROM, and squatting requires at least 130° ROM [2]. Stiff knees cause a variable degree of disability and interferes with activities of daily living. Stiff knees can occur due to a multitude of conditions like inflammatory arthritis, advanced degenerative arthritis, post-traumatic arthritis, sequelae of previous infection, healed tuberculosis, and in rare cases reflex sympathetic dystrophy and hemophilic arthropathy. Total knee arthroplasty (TKA) in these stiff knees is particularly challenging. The pathoanatomy in stiffness/ankylosed knees in flexion and stiffness/ankylosed knees in extension is different. The major problems encountered in stiffness in extension are adhesions in suprapatellar pouch and adhesions in the medial and lateral parapatellar gutters, contracture of extensor mechanism, patellofemoral joint fusion, loss of tibiofemoral joint space, difficult dislocation, jog of a motion, or fusion in extension. Stiffness in flexion usually needs addressing upon the flexion contracture and posterior releases as well as releasing intraarticular adhesions and posterior osteophytes that cause a bony block. Stiffness in flexion is different, because here the quadriceps is not contracted, and knee can be subluxed and so it is easier to obtain flexion after a TKA. Ankylosed knees on the other hand have no motion at all with no joint line differentiation possible. The collaterals and capsule are contracted and cruciates are usually absent. Extensor mechanism may be severely contracted in ankylosis in extension or may be preserved in flexion. The variable pathoanatomy of stiff/ankylosed knees in flexion and extension requires specialized exposure, need to address varied degree of bone loss, and instability may require special constraints/hinged implants. Nevertheless, the results of TKA in stiff knees are inferior to routine TKA with multiple factors coming into play like the degree of soft-tissue contractures, bony/fibrous ankylosis, preoperative range of motion, patient motivation, and compliance and surgeons’ expertise [3].

In our article, we have tried to classify the stiff knees based on our experience of managing these complex knees. This classification would give the surgeons a direction on how to approach a stiff knee, which implants to choose and what can be the predicted outcome in such situations. The classification has a prognostic value and helps the surgeons explain to the patients about the outcomes during preoperative counseling.

Materials and Methods

It was a retrospective study conducted in our institute to study the outcomes of TKA in stiff/ankylosed knees and design a working classification based on the deformity, ROM, exposure, prosthesis used, and functional outcomes seen in our series of patients.

Of the 570 TKA performed in last 5 years, 69 had stiffness. Eight were lost to follow-up. Out of the remaining 61 knees, 57 knees included in the study and had stiffness due to fibrosis and four had bony ankylosis. 26 were men and rest were females. Patients were aged between 39 and 76 years (av. 56.2 years). The main reasons for stiffness/ankylosis were advanced degenerative arthritis (n = 27), inflammatory arthritis (n = 21), post-traumatic arthritis (n = 6), sequelae septic (n = 4) and healed tuberculosis (n = 2), and arthrodesis (n = 1). Preoperative AP and lateral radiographs of knee joint were done to assess the alignment and clinical ROM was calculated. Forty nine has a varus deformation av. 8° (range 3°–17°) and rest 12 had a valgus deformation av. 6° (range 3°–11°). Knee society scores (KSS) were recorded preoperatively and postoperatively at 6 weeks, 3 months, 6 months, and then yearly. Knees were classified into three groups depending upon whether they had fibrous or bony ankylosis and further subdivided based on the preoperative ROM (Table 1). All patients were operated under combined spinal epidural and under tourniquet by the same surgeon (MS).

Table 1.

Classification for stiff/ankylosed knees

Type 1:Stiffness/fibrous ankylosis in flexion

•1a:Stiffness/fibrous ankylosis in 30°–60°

•1b:Stiffness/fibrous ankylosis in 60°–90°

Type 2:Stiffness/fibrous ankylosis in extension

•2a:Stiffness/fibrous ankylosis in 0°–15°

•2b:Stiffness/fibrous ankylosis in 15°–30°

Type 3:Bony ankylosis/arthrodesis

•3a:Bony ankylosis in flexion > 30

•3b:Bony ankylosis in extension (flexion < 30)

Surgical Considerations

A standard medial parapatellar approach is used to begin with. For knees where a difficulty in exposure is encountered, an extensile exposure like quadriceps snip (QS) or tibial tubercle osteotomy (TTO) along with release of adhesions was performed. A standard medial parapatellar approach was used in 42 knees, a quadriceps snip was used in 16 knees and a TTO was used in three knees. After the suprapatellar area and gutters were cleared and joint synechia excised, the knee could be further flexed and a distal femur cut performed first (possible only if the knee has been flexed to 45° and then only can a femoral intramedullary rod be placed). Forceful flexion should be avoided as it can lead to a fracture of posterior femoral condyle or a patellar tendon avulsion. In bony ankylosed knees or where the knee could not flexed to 45°, an in situ tibial cut is recommended. Alternatively, a TTO may be performed. A bony ankylosis would require a patellofemoral osteotomy, release of gutters, and parapatellar area like a femoral peel followed by an insitu tibial cut where a biscuit of bone was removed to create joint space. This is followed by routine steps.

A posterior stabilized (PS) implant sufficed in 42 knees, and varus–valgus constrained (VVC) implants were used where stability could not be achieved with standard cruciate substituting implants in 12 knees. A hinged prosthesis was used where there was a greater flexion extension mismatch and mediolateral instability in four knees and a megaprosthesis was used in three knees.

Table 2 shows the outcomes of patients according to the classification.

Table 2.

Class-wise knee society scores and range of motion

Classification	Exposure	Bony cut	Implant	Prognosis KSS and ROM (°)
Type 1a N = 12	Parapatellar with extension to Quadriceps snip	Distal femur first	PS/VVC	GOOD KSS > 80 ROM 90°–110°
Type 1b N = 27	Parapatellar exposure	Joint can be subluxed	PS/VVC	EXCELLENT KSS > 90 ROM > 110
Type 2a N = 7	Quadriceps snip	In situ tibial cut	VVC	AVERAGE KSS < 60 ROM 60–80
Type 2b N = 11	Quadriceps snip	Distal femur cut	PS/VVC	AVERAGE–GOOD KSS 60–80 ROM 80–100
Type 3a N = 3	Quadriceps snip/TTO	In situ tibial cut	Hinged implants	Poor KSS < 50 Extension lag ROM 60–80
Type 3b N = 1	Tibial tubercle osteotomy/quadricep snip/pie crusting Rectus	In situ tibial cut	Hinged implants/megaprosthesis	Poor(< 50) Extension lag, ROM < 60 Limited flexion

Type 1a: Stiffness/Fibrous Ankylosis in 30–60 (n = 12)

Exposure usually required a standard approach which could be extended to a quadriceps snip.

Joint could be flexed further after releases in gutters. The distal femur cut was performed first and posterior stabilized implants sufficed (keeping VVC as standby, (Fig. 1).

Fig. 1.

a Preoperative AP and lateral views of a 69 year old female with advanced osteoarthritis and stiffness. b Preoperative weight bearing clinical picture of the patient showing bilateral severe varus. c Clinical picture showing the fixed flexion deformity of 45°. d Intraoperative picture showing the completely eroded cartilage and bone defect on medial proximal tibia. e Postoperative Ap radiographs of the knee showing the corrected alignment. f Postoperative lateral radiographs of the knee joint showing the correction of the FFD. g Follow up X-rays of same patient at 2 years

Type 1b: Stiffness/Fibrous Ankylosis Beyond 60° (n = 27)

Exposure required a routine medial parapatellar exposure and tibial could be subluxed and a tibial cut could be performed first. These required extensive posterior capsular and osteophyte release. Enhanced distal femoral resection was required to create extension gap. Posterior substituting implants suffice, but VVC may be required (Fig. 2).

Fig. 2.

a Preoperative AP and lateral views of a 39 year male with h/o prior knee sepsis 17 years ago. He presented with an FFD of 60° and shortening of 4 cm. b Intraoperative picture showing the intraarticular synechia. c Postoperative AP view of the knee joint with constrained VVC implants and stem extenders in both femur and tibia. d Postoperative lateral view of the same patient. e, f Same patient developed infection and was treated with a DAIR. Antibiotic beads (Stimulan) can be seen in the knee joint. Infection healed at follow-up

Type 2a: Stiffness/Fibrous Ankylosis in 0°–15° extension (n = 7)

Patellofemoral joint is commonly adhered or fused and may require a patellofemoral osteotomy. Exposure requires a quadriceps snip. Tibial tubercle osteotomy is rarely needed, but is recommended if a patella baja is present. In situ tibial cut with preservation of collaterals is required in almost all the cases as the knee cannot be dislocated. May requires VVC implants because of flexion extension mismatch (Fig. 3).

Fig. 3.

a Preoperative Ap and lateral views of the knee joint in a 55 year female suffering from advanced stage rheumatoid arthritis with fibrous ankylosis in extension. Note the patellofemoral joint fusion seen on left side. b Intraoperative picture of right knee showing the release of parapatellar gutters and suprapatellar area. c Intraoperative picture showing the in situ tibial cut being performed, since the knee cannot be subluxed. d Intraoperative picture showing the removal of a biscuit of bone to create joint space. e Intraoperative picture showing that the knee joint can be bent to 45° and an intramedullary rod can be introduced to perform distal femur cut. f Intraoperative picture showing the sizing jig placed to determine the size and femoral component rotation. g Intraoperative picture showing the final implants. h Postoperative Ap X-rays of the same patient with a PS implant. i Postoperative lateral views of the knee joint

Type 2b: Stiffness/Fibrous Ankylosis in 15°–30° (n = 11).

Patients usually have a jog of motion. Exposure requires quadriceps snip, and after clearing the gutters, flexion can be achieved till 45°. This is when a distal femoral cut can be performed. Patient may require PS or VVC implants (Fig. 4).

Fig. 4.

a Preoperative Ap and lateral views of a 41 year female suffering from rheumatoid arthritis with fibrous ankylosis in extension and only a jog of motion present. She was operated 5 years ago. b Clinical picture taken in operating room showing the Knee stiff in extension (20°). c Postoperative Ap and lateral views with stem extenders used in Tibia. d clinical picture in immediate postoperative period showing the knee bending to 100°. e 4 years after the index surgery, she falls and sustains a periprosthetic fracture femur. F Fixation done with a locking plate. Prosthesis was well fixed. g, h Postoperative flexion achieved till 90° and extension lag was seen in left knee which improved over time

Type 3a: Bony Ankylosis in Flexion > 30° (n = 3)

Exposure requires a quadriceps snip or a TTO. Tibial cut needs to be performed in situ as the knee cannot be dislocated. Hinged implants need to be used (Fig. 5).

Fig. 5.

a Preoperative X-rays of knee joint in a 21 year male showing a bony ankylosis in flexion of the right knee resulting from a healed tubercular infection 8 years ago. B Clinical picture showing the knee joint in flexion and apparent shortening. c Intraoperative picture showing the patellofemoral osteotomy being performed. d Distal femur cut is followed by an in situ tibial cut to create joint space. Surgeon has to be cautious in preventing accidental injury to the blood vessels behind the joint. e Intraoperative picture showing the S-rom Knowles hinge being implanted. f Two year follow-up X-rays of the same patient showing the metaphyseal sleeves in the femur and a hinge prosthesis. g Clinical picture of the same patient showing full extension, a flexion of around 90°, and an extension lag. However, he is walking comfortably and pain free with full weight bearing on the operating leg

Type 3b: Bony Ankylosis in Extension < 30° (n = 1)

A poorly performed arthrodesis is an indication to knock down the fusion and convert to a functional knee. The approach is complex and may require a Quadriceps snip or a TTO or a pie crusting of the rectus is recommended. The authors believe that VY plasty is a debilitating approach with poor results. This category always requires hinged implants, because the collaterals are absent and there is a huge flexion–extension mismatch. A hinged megaprosthesis was indicated if the bone loss was extensive (Fig. 6).

Fig. 6.

a, b Preoperative Ap and lateral views of knee joint of a 50 year male with failed arthrodesis (index surgery done 8 years ago) for infection and a broken Kuntscher nail. c After infection was ruled out and keeping in mind the bone loss, we performed replacement using a tumor megaprosthesis. Ap view of the knee shows that the remanent of K-nail was left as such as we were not able to remove the same and it did not interfere with the placement of the femoral stem into the canal. d Postoperative lateral view of the knee and femur showing the placement of megaprosthesis

Results

Patients were followed for an average 2.4 years (1.8–5.5 years). KSS pain scores improved from an av. 32 preop (18–64) to av. 76 postoperatively (61–90). The KSS function scores improved from a preoperative value of 36 (16–56) to an av.78 (52–90) postoperatively. ROM improved from an average of 35.6° (0°–44°) preoperatively to an average of 95.6° (ROM 73°–118°) postoperatively, an average gain in arc of motion of 60°. Extension lag was an av. 8° (3°–12°) and seen in 13 patients postoperatively. For knees fixed in extension for a long time, the residual quadriceps muscle fibers recruited over a period of 3 months and the extension lag was corrected to an extent. Residual fixed flexion deformity was an av. 7° (3°–14°) and was seen in 17 patients. This was commonly seen in rheumatoid patients who were put on a push lock knee splint for around 6 weeks and this resolved to an extent. The class-wise results are depicted in Table 2. Type 1 (n = 39) with fibrous ankylosis in flexion needed routine exposure and PS implants and achieved good KSS > 80 and ROM > 100. Functional prognosis was good and superior to type 2. Type 2 (n = 18)with fibrosis had average KSS 60–80 and ROM 80–100 but comparatively poor outcomes as compared to the type 1. Type 3 (n = 4) bony ankylosis achieved an ROM < 80. KSS < 60. The stiff knees (type 1 and type 2) fared better than ankylosed knees (type 3) in all aspects.

The complication rate was high (24%) in our series with skin necrosis seen in three patients, residual stiffness in six knees, hematoma formation in six knees, heterotrophic ossification in two knees, and infection in one patient. Patellar tendon rupture occurred in one patient intraoperatively for which an end-to-end repair and augmentation was done with semitendinous graft. Infection was managed with a DAIR (debridement, antibiotics, polyethylene exchange, and implant retention) (Fig. 2). Common peroneal palsy was not found in our series. Periprosthetic fracture was seen in one patient after 3 years who fell off the stairs. This was stabilized with a locking plate (Fig. 4).

Discussion

Stiff knees pose considerable challenge to the operating surgeon in terms of exposure, soft-tissue balancing, component selection and placement, and outcomes are not desirable despite the best effort and surgical skills. Complication rate is high and functional outcomes are inferior.

Exposure

The authors prefer a medial parapatellar arthrotomy with an extension into a quadriceps snip if needed. Quadriceps snip is easier to perform and does not alter the rehabilitation protocol or the functional outcomes as compare to the standard approach [4–6]. V–Y quadriceps plasty although provides an extensile exposure, it commonly results in an extension lag and modification of the rehabilitation protocol [2]. The senior author (MS) has never used a V–Y quadriceps plasty in the series as he believes that it is too much debilitating, needs bracing, delays rehabilitation, and leads to residual extension lag. Kim et al. used this approach in their series of 99 ankylosed knees, whereas Rajgopal et al. and Bhan et al. preferred the quadriceps snip [3, 7, 8]. Tibial tubercle osteotomy (TTO) is usually recommended in stiffness/ankylosis in extension or in patients with patella baja, because it helps in increasing the quadriceps length by proximal migration of patellar tendon attachment and hence gaining excursion of the extensor mechanism which increases the possible flexion. The problems mainly seen are difficulty in kneeling, hardware-related problems, and extension lag (although it is less as compared to the V–Y quadriceps plasty) [6]. Cameron et al. recommended using a TTO for all knees ankylosed in extension as it lengthened the extensor mechanism [9].

Prosthesis Selection

Stiff and ankylosed knees are complex and the surgeon should keep all the constraints ready in their armamentarium including the hinged knee prosthesis and even a megaprosthesis. The authors could manage to implant a posterior stabilized prosthesis in majority of the knees in type 1 and type 2, because the stiffness of the soft-tissue sleeve around the knee provides an inherent stability. Moreover, the PCL is usually deficient in these stiff knees, so the surgeon should be vary of using a cruciate retaining implant. Type 2 with major flexion extension mismatch required VVC implants and type 3 almost always required a hinged prosthesis or a tumor prosthesis. Whereas Bhan et al. supported the use of a constrained prosthesis in stiff knees, Aglietti et al. recommended using PS implants quoting that stiff knees require more soft-tissue releases [2, 3]. Contrary to this Fosco et al. advocated using unconstrained implants, stating that minimal and titrated soft-tissue releases should be performed [10].

Functional Outcome

No matter what approach is chosen or what implants are used, the ROM in a stiff knee is inferior to that achieved in a flexible knee [11]. The causative factor for stiffness has been shown to be a deciding factor in attaining the postoperative ROM [11, 12]. Our results also corroborate with the findings in the literature which shows inferior results for Tkr in stiff/ankylosed knees as compared to normal knees. Kim et al. reported on 86 tkr in stiff knees with a mean preoperative ROM of 40 (range 10°–50°). The Hospital for Special Surgery Knee score, KSS, and FS scores improved from a preoperative 42, 11, and 42 points, to a postoperative 84, 90, and 84 points, respectively. They reported a complication rate of 14% [13]. Rajgopal et al. in a recent article on 115 stiff knees with an ROM between 0° and 20° reported a complications rate of 20.9% (24 knees). Nine knees had an extension lag (10°–20°), and others showed stiffness, flexion contracture, infection, skin necrosis, component loosening, peroneal nerve palsy, and heterotrophic ossification [14]. Bhan et al. reported better results in patients with stiff knees than in patients with ankylosed knees in a series of 90 TKA (52 patients) at minimum 2 year follow-up (average of 6.5 years). They found better KSS scores and arc of motion for stiff knees. Major complications were seen in 4 of the 26 ankylosed knees and in 2 of the 64 stiff knees [3]. Young et al. reported a higher complication rate of 33.3% in their series on painless stiff knees. They reported 15 knees followed for a period of 15 years with a preop ROM 0°, fusion angle of 11.6 ± 17.9° (range 0°–50°) in a flexion position. Seven knees had a knee arthrodesis. Quadriceps turn down was performed in 13 and TTO in 2 cases. VVC implants were used in 12, PS in 3. Total complication rate was 33.3% including a tibial tubercle avulsion, infection, and patellar fracture. Extension lag seen was 10°–15° in their series [15]. Kim et al. reported lower HSS scores and flexion of 77.7° in a group of 36 ankylosed knees followed for a period of 7.7 years. They reported a high skin complication rate of 50%, quadriceps rupture in two patients and deep infection in another two. The contradictory finding was their recommendation to use a PS implant in fused knees, stating that the tight soft-tissue compensates for the deficient collaterals [16]. Rajgopal et al. have also reported in a series of 84 stiff knees with 4–15 year follow-up and reported a postoperative arc of 75° with a residual flexion contracture of 5° and a gain in motion of av 61° [8]. We have had similar outcomes in our series with improvement in KSS scores, and gain in average 60° arc of motion and a flexion deformity of 7°.

Kernkamp in a meta-analysis including six papers and 123 knees converted from arthrodesis into arthroplasty, found a very high complication rate of 65% including a skin necrosis in 25% patients, infection (11%), arthrofibrosis (13%), revision (11%), and major complications like amputation and death in less than 5% [17]. Jaureguie in another meta-analysis on conversion of 98 arthrodesed knees to TKA and followed for 5 years concluded that it is a challenging procedure to be done by expert surgeons and the results are satisfactory but reported a high complication rate of 47% [18]. Rai et al. reported outcomes of TKA in 38 patients with stiff knees post-trauma operated using constrained implants and followed for 6.47 years. They reported an improvement of KSS pain scores from 44 preoperatively to 91 postoperatively and an improvement of KSS function score from 44 preoperatively to 91 postoperatively with a 94% implant survival at last follow-up [19]. Bhan et al. have compared the outcomes in stiff (n = 64) and ankylosed knees (n = 26), and reported better outcomes and less complications in stiff knees (2/64) compared to 4/26 in ankylosed knee [3]. We also have had similar findings with bony ankylosed knees (type 3) showing the least improvement in KSS and ROM and a relatively poor function (Table 2). Rajgopal et al. have compared the outcomes of TKA in 51 knees stiff in extension with 64 knees stiff in flexion followed for over a period of 9.7 years. They concluded that the flexion group had a better arc of motion but also reported more residual flexion contracture. They reported similar functional scores and complication rate in both the groups. We had better functional outcomes, better ROM, and lesser complications in the flexion group (type 1) as compared to the extension group (type 2), but the least favorable outcome was seen in the bony ankylosis group (type 3) [14].

Complications

Majority of the studies available in the literature on stiff knees have reported a very high complication rate ranging from 14 to 65% [3, 8, 14–19]. Skin necrosis is rampant in stiff knees because of the poor vascularity due to multiple previous surgeries, scarring, and stiff tissues around the knee joint. We noticed relatively less skin necrosis (three knees) as compared to other studies which we believe depends upon the soft-tissue handling. Some authors have recommended using tissue expanders before surgery, but we have not used any in our series [20, 21]. Common peroneal palsy is a reported complication especially in the ankylosed knee in flexion and can occur due to overzealous correction of deformity [22]. Montegomery et al. have reported two cases of peroneal nerve palsy in their series of 82 arthroplasties for stiff knees [23]. No case of peroneal nerve palsy was seen in our series. Debette et al. in a large series of 304 stiff knees (ROM < 90°) reported a 17% complication rate with pain and residual stiffness as major complications [24]. We had residual stiffness in six cases which were managed conservatively with physiotherapy only. Component revision as an end point was not seen in our series, although a revision rate of 4% has been reported by Kim et al. and 17% by McAuley et al. [7, 25]. High revision rates up to 35% have been reported by Naranja et al. in their study on TKA in arthrodesed knees [26].

Infection can be a dreaded complication and preexisting comorbidities, increased duration of surgery, and soft-tissue handling have a bearing on the incidence. Although hematoma formation is common occurrence in stiff knees due to the extent of soft-tissue releases, deep infection was seen in only one patient in our series. This was revised with a DAIR procedure with no recurrence of infection. Kim et al. have reported two pyogenic infections in 36 stiff knees [16]. Kernkamp in their meta-analysis on conversion of arthrodesed knees reported an 11% rate of infection [17]. Infection rate has been more or less consistent in all studies on stiff knees.

Quadriceps tendon or patellar tendon rupture can occur during forceful exposure. Kim et al. have reported two cases of quadriceps tendon rupture [16]. We did not see any quadriceps rupture because of our low threshold to use a quadriceps snip in stiff knees. Naranja et al. in their series on 37 stiff knees have reported three cases of patellar tendon/tubercle avulsion [26]. We had a patient with patellar tendon rupture who had a patella baja with stiff knee in extension and thinned out and adhered patellar tendon. The author recommends to peel off the patellar tendon from the proximal tibial to relax the pull on the tendon, keeping intact the attachment to the tibial tubercle. A prophylactic pin placed into the tibial tuberosity also prevents tibial tubercle avulsion. Stiff knees can develop a multitude of complications if prophylaxis is not taken during the perioperative period.

Our study on stiff knees is the first to define a working classification which has a prognostic value. The major limitation of our study is that the number of patients in the study is small and from a single surgeon and single center. We would recommend orthopedic surgeons to use the classification in their practice, so that multicentric studies with larger groups can be published.

Conclusion

TKA in stiff/ankylosed knees is a difficult preposition. Standard medial parapatellar approach with or without a quadriceps snip suffices exposure in 90% of cases and causes minimal extension lag. Instability is not an issue even with PS implants, but VVC constrained implants should be kept handy (type 2a, b: flexion extension mismatch). Results are generally poor in bony ankylosis (type 3a, b). Stiff/ankylosed knees have a high risk of complications. This classification of stiff/ankylosed knees guides the surgeon to decide upon which approach to take, which implants to keep handy and has a predictive and prognostic value.

Declarations

Conflict of Interest

The corresponding author on behalf of all the authors declares that they have no conflict of interest to declare and no source of funding has been received for this study.

Ethical Approval

This article does not contain any studies with human or animal subjects performed by the any of the authors.

Informed Consent

The authors would like to state that all patients were operated after informed consent and the study was done after approval of the internal review board.